Fig. 1.1
Patient with right ankle injury treated with open reduction and internal fixation of fibula and closed treatment of distal tibia fracture. Referred for nonunion after progressive deformity developed. a–c Three views (anteroposterior [AP], mortise, lateral) of the right ankle show failure of the fibula hardware and mal-alignment with nonunion of both the tibia and fibula. Patient underwent hardware removal and cultures. d–f Three views (AP, lateral, and mortise) of the right ankle after hardware removal. Due to the malalignment and stiff nonunion, a Taylor spatial frame (TSF) was applied to allow correction and healing of the nonunion. g–h AP and lateral after TSF applied to right ankle prior to correction. i–j AP and lateral with TSF showing full correction of the deformity and realignment of the limb. k–m Three views (AP, lateral, and mortise) of the right ankle 1 year after consolidation of nonunion and removal of TSF
Patient medical factors contributing to a biological cause for the nonunion are many and can be problematic not only from the original fracture standpoint but also for the treatment of an established nonunion [10, 13, 25, 32, 40]. Established diseases such as vascular disease, rheumatologic disease, and s/p organ transplantation cannot be affected, but their effects on fracture healing and subsequent management of the nonunion need to be taken into consideration. Perhaps their steroids or immunosuppressive agents can be held for short time period which would allow for surgical intervention and healing, and such decisions should be made in conjunction with the patients’ appropriate other physicians. A multidisciplinary approach is needed to get many of these patients healed.
Although there are many endocrine abnormalities that can affect the musculoskeletal system, such as thyroid and parathyroid disorders, hypogonadism, and calcium imbalances to name a few [13], diabetes has had the most attention due to the high prevalence in the population. Diabetes has been shown to prolong healing times for fractures [40, 41]. It is also well documented that patients with diabetes have increased complications when dealing with musculoskeletal conditions, especially with fractures [32, 42, 43]. In a nationwide population based study out of Taiwan, diabetics were found to have an increased incidence of fractures as well as more adverse events and a higher mortality after fractures [42]. The addition of neuropathic complications can make even simple fractures that require surgery end up being disastrous for the patient. Wukich et al. [43] showed that patients with ankle fractures that had complicated diabetes had a 3.8 times increased risk of overall complications and a 3.4 times increased risk of malunion and nonunion compared to uncomplicated diabetic patients. These patients were also 5 times more likely to require revision surgery or arthrodesis. Diabetics need to understand that glucose control is extremely important for them to avoid diabetic complications of end organ damage, neuropathy, nephropathy, and peripheral arterial disease to minimize further musculoskeletal complications [32]. Diabetics should be treated with prolonged immobilization and delayed weight bearing compared to the nondiabetic to aid in avoiding complications. Additionally, many of these patients require additional fixation for otherwise straightforward fractures to try and prevent the late complications that occur with these injuries.
Vitamin D deficiency or insufficiency has been linked to nonunions, but a clear causal link is difficult to establish [13, 40]. Both the 25-OH vitamin D and 1, 25 OH2 vitamin D levels can be monitored, but the 25-OH level is the one that is important. Patients with 25-OH levels <20 are considered insufficient and between 20 and 30 deficient. It is not clear however whether higher levels than simply above the 30 level are needed in patients with fractures. Brinker et al. [13] showed that a preponderance of their nonunion patients had vitamin D deficiency. They had 37 patients that were evaluated for a metabolic or endocrine abnormality of which 68% (25 of 37) had vitamin D deficiency . It has become increasingly clear that many patients are vitamin D deficient or insufficient. In a meta-analysis of the literature, it was found that the pooled prevalence of hypovitaminosis was 77.5% in young trauma patients and 73% in geriatric fragility fracture patients [44]. In a follow-up study, the same authors showed that there is a lack of consensus in prescribing vitamin D to fracture patients. They found that 66% of surgeons tended to prescribe vitamin D to fragility fracture patients compared to 25.7% to nonfragility fracture patients [45]. The lack of prescribing in this population needs to be re-examined since the prevalence of low vitamin D in young trauma patients is high. Low vitamin D is more prevalent than previously thought and is widespread in patients of all orthopedic subspecialties and not just orthopedic trauma [46]. Management of vitamin D is easily done via replacement therapy and has been shown to be successful in raising serum levels [44]. In a study to evaluate the cost benefit of both calcium and vitamin D supplementation in all fracture patients, the cost of an 8-week course of treatment was determined and compared to the cost savings assuming just a 5% reduction in nonunions. This would result in a potential cost savings of $65,866 annually [47]. Many dosages of replacement therapy are available, but the authors’ preference is for high-dose (50,000 IU) vitamin D weekly for six months along with calcium supplementation. The target is to obtain a 25-OH level in the 40–60 range. Patients with low vitamin D can also develop secondary hyperparathyroidism and should also have a parathyroid hormone (PTH) level drawn when evaluating for a nonunion. The high PTH can contribute to the development of a nonunion [13]. In most cases, the high PTH will resolve with appropriate vitamin D replacement therapy.
Osteoporosis has also been linked to the development of nonunions [32]. The issues with osteoporotic bone healing are both biologic and mechanical [48]. By definition, osteoporotic bone is bone with less bone mass and as such is at an increased risk for fracture. The biologic changes that occur with osteoporosis, including a diminished level of mesenchymal stem cells and thus osteoblasts, a decrease in the chondrogenic potential of the periosteum and other alterations in the fracture healing pathway results in a less than robust fracture healing process [32, 40, 48]. Additionally, because of the lower bone mass, the fixation in such bone can be problematic and as such can lead to inadequate fixation and fracture stability. The result can be a nonunion. Many specialized techniques have been described in the management of osteoporotic fractures and should be employed when dealing with nonunions especially if mechanical failure was a significant contributor to the development of the nonunion. Locked plating, use of load-sharing devices, use of fixed angle devices, augmentation of fixation with cement or bone graft substitutes, adjunctive use of structural bone grafts, and preservation of soft tissue can assist in the management of these fractures and nonunions [48]. Although most osteoporotic individuals are elderly, age is an independent factor which can negatively affect fracture healing also resulting in delayed unions or nonunions [49]. This decline in healing potential can be attributed to hormonal changes, changes at the cellular level of fracture healing signaling, and diminished mesenchymal stem cells which all may also occur with osteoporosis. The true etiology still requires much more investigation due to the complex interplay that occurs in fracture healing and the overlap in physiology with aging and osteoporosis. Another confounding factor is that patients with osteoporosis are often being pharmacologically treated for their osteoporosis. The most common are the bisphosphonates , which are anti-resorptive agents and inhibit osteoclast function. The interference with remodeling of the bone has resulted in an unwanted side effect and resultant “atypical” femoral fractures. It is advised that these medications be discontinued during the fracture healing process [32, 48]. It is also important to evaluate all their medications and the potential effects that they may have on bone metabolism.
One of the most common class of medications that many patients take, both prescription and over the counter, are the nonsteroidal anti-inflammatory drugs (NSAIDs) . Recently, the use of NSAIDs during fracture healing has come under intense scrutiny [49–51]. Early reports of the use of NSAIDs in animal fracture healing models showed a clear deleterious effect [32, 49–51]. The doses required were very high. The mechanisms by which they are theorized to inhibit fracture healing include inhibition of prostaglandin synthesis and reduction of osteoblast activity, both of which result in an impaired fracture healing response [32]. Prostaglandins are needed during the inflammatory phase of fracture healing and help start the osteogenic response [6, 31, 38, 49]. Although a few clinical studies have shown a loose association between the use of NSAIDs and nonunions, it is controversial [50]. Kurmis et al. [51] performed a systematic analysis of over 300 relevant papers and concluded that there was not significant evidence to indicate a negative effect on fracture healing from the short-term use of NSAIDs after a fracture. Most of the clinical studies published in relation to NSAIDs in fracture healing were Level 5 evidence or expert opinion only [28]. Therefore, it is hard to make a recommendation on the use of NSAIDs both in terms of timing and duration immediately after a fracture. Due to the lack of guidelines and unknown true effects on fracture healing, the author’s practice is to avoid NSAIDS for the first 4–6 weeks after a fracture. This is especially true of Indomethacin. Additionally, we do not use Toradol (intravenous or per os) intra-operatively or immediately postoperatively for acute fracture cases. NSAID use after repair of nonunions has not been investigated to our knowledge.
Since inflammation is one of the initiating factors for bone healing, it has been suggested that perhaps healing may be altered in the polytrauma patient as well [32, 37, 40, 50]. These patients undergo a prolonged state of inflammation [40]. It is thought that the increased inflammation could delay fracture healing through a variety of cellular responses [20, 50]. Additionally, many of these patients also have multiple fractures that may require operative intervention. The post-op rehab protocol on one fracture may result in delayed stimulation of another fracture with resultant delayed healing or even nonunion. Other system injuries also may have an effect on fracture healing as well. A literature review by Hildebrand et al. [37] found that isolated hemorrhagic shock, chest trauma, severe soft tissue injury, and systemic inflammation can all affect fracture healing. Finally, it has been suggested that the American Society of Anesthesiologist (ASA) classification, which indicates overall health, was associated with nonunion development—the higher the ASA, the increase in probability of a nonunion [23].
Smoking has been clearly shown to inhibit fracture healing and result in both delayed unions and nonunions as well as increase the overall complications in the management of fractures [10, 21, 24, 28, 29, 32, 46, 49, 52–54]. Smoking has also been linked to an increase in fracture rates of the hip, distal radius, spine, and other osteoporotic fractures [52]. The exact mechanism and offending agent has not been clearly elucidated. Nicotine is one of more than 4000 chemicals that exist in cigarette smoke [46]. It has been shown to cause vasoconstriction (resultant hypoxia), platelet adhesion, and reduced cell proliferation for healing. All of these physiologic changes result in a negative effect on both wound and fracture healing [52]. However, it is not clear exactly which chemical is responsible for all the negative effects. Animal studies have had conflicting results with some studies, with nicotine alone, not showing the negative effects that are seen with smoking, whereas others have shown deleterious effects [49]. Nevertheless, the clinical literature overwhelmingly supports the increased risk of delayed unions, nonunions, and wound complications seen with smoking [21, 24, 52–54]. Scolaro et al. [54], in a systematic review of the literature, showed that smokers were over 2 times more likely (statistically significant) to develop a nonunion than nonsmokers. This was especially true in open fractures and tibia fractures. There was also a trend for longer healing times and infections (deep and superficial) in the smoking group. In a separate systematic review of the literature, Patel et al. [53] also found a negative effect of smoking on bone healing. They also looked at each study in relation to the bone or procedure in question. All the tibia fracture studies, except for one treated with external fixation, showed a clear increased risk of nonunion from smoking. This was also true in distraction osteogenesis, fibula fractures, ulna osteotomy healing, subtalar and ankle arthrodesis, and elective foot surgery. Fractures of the femoral diaphysis were not statistically significantly affected by smoking. In contrast, Hernigou and Schuind [21], in their retrospective study looking at diaphyseal fractures, found that smoking was significantly associated with nonunions (OR 8.25) in the femur, as well as the tibia and humerus. Westgeest et al. [29] found that in a prospective cohort study of open long-bone fractures, smoking (OR 1.73) was significantly associated with developing a nonunion. Murray et al. [24] looked at their series of diaphyseal clavicular fractures. They found that smoking was the strongest predictor of a nonunion (OR 3.76) and recommended that smoking cessation be an integral part of any treatment. However, getting patients to stop smoking is extremely difficult. The first step is acknowledging that smoking is bad for one’s health. Matuszewski et al. [55] performed a cross-sectional cohort survey study and found that smokers did not understand the negative effects of smoking on their general health or on fracture care. On a positive note, the orthopedic trauma patients surveyed seemed interested in smoking cessation more so than what was expected. They recommended formal education for smoking cessation. It is well accepted and has been shown that preoperative smoking cessation can reduce both pulmonary and wound complications postoperatively [46]. Educating the patients on the ill effects of smoking on fracture healing is part of our “discussion” with the patient being evaluated for nonunions. It is the author’s policy to not perform nonunion surgery on active smokers as long as the management can be done on an elective basis (aseptic nonunions). Both serum and urinary levels of cotine and nicotine are monitored to insure patient compliance. Although many feel that smoking cessation is the primary care physician’s responsibility, as an orthopedist it behooves us to play an active role to help maximize the patient’s outcome and minimize complications from any surgical intervention.
When evaluating a patient for a nonunion management, one must assess for the presence of the risk factors above. There are certainly more comorbidities than can affect fracture healing, but these are the most prevalent. These risk factors and/or co-morbidities should be improved upon or corrected if feasible. Many are injury or treatment related, but knowing those details can help devise an appropriate treatment plan for the nonunion.
1.4 General Principles
1.4.1 Diagnosis
The diagnosis of a nonunion is highly controversial because no gold standard exists for healing assessment [6, 15]. In a multinational survey of orthopedic surgeons, there was a 73 and 53% consensus that a lack of standardization in the definition for a delayed union and nonunion, respectively, existed. However, they did agree (88%) that the diagnosis should be done based on clinical evaluation and plain radiographs [3]. Pain on weight bearing was felt to be the most consistent predictor of delayed union and nonunion.
The diagnosis of a nonunion should be made on a series of radiographs in addition to the clinical picture. Often the fracture healing may be delayed, but critical evaluation of radiographs 6–8 weeks apart may show some improvement indicating progress. If the X-rays show no progress on two sets of consecutive images and the patient is having pain, then nonunion has probably been established assuming sufficient time has initially passed. The problem arises in the patient without symptoms but clear radiographic evidence of a nonunion. Many of these patients, because they lack symptoms, may not return in fear of needing surgery. The problem occurs when they return after hardware failure with new-onset pain and/or deformity. The time passed based on the FDA definition may not have been reached, but if cessation of all healing is indicated by plain radiographs and the patient is symptomatic, then intervention is probably warranted.
1.4.2 Radiographic Evaluation and Scoring
After the history and physical, evaluation should always begin with plain radiographs. It still remains the most common method of assessing for fracture union. However, just as in the lack of standardization of definitions, there is a lack of consensus on radiographic criteria as well. Dijkman et al. [56] reviewed the literature to look at radiographic criteria used in studies. They found that bridging of the fracture by bone, callus, or trabeculae was used 53% of the time. Bridging of the fracture across three cortices 27% of the time and loss of fracture lines was 18% of the time. The best interobserver reliability was found to be the number of cortices bridged by callus.
Despite the issues with radiological criteria , standard orthogonal views (anteroposterior and lateral) of the bone in question should be obtained. If the patient is referred in, previous studies are desired for comparison. In some cases, the fracture is actually progressing and reassurance is all that is needed. They may have a delayed union, but radiographic evidence of healing is occurring. The length, alignment, and rotation of the limb should be appropriately evaluated. In the lower extremity, if there is an associated deformity, then additional full-length radiographs (± ruler) from the hip to the knee are obtained to assess the mechanical axis of the limb (Fig. 1.2). Restoring the mechanical axis of the limb can aid in healing of the nonunion and should be part of the preoperative plan. If it appears to be short, then a scannogram (Fig. 1.3) or full-length radiographs with a ruler should be obtained (see Fig. 1.2 ). Oblique radiographs can aid in the diagnosis as well, if the standard anteroposterior and lateral do not clearly show the nonunion due to the obliquity of the original fracture or because of overlying hardware (internal or external fixation). Such views can also better define the plane of maximum deformity when that plane is not in the usual sagittal or coronal plane. Rotation can be assessed clinically in some situations; otherwise, a CT scan may be needed (Fig. 1.4).
Fig. 1.2
Patient with a long leg film obtained for evaluation of his mechanical axis with a left femoral neck nonunion . A ruler can be used also to evaluate for leg length discrepancy
Fig. 1.3
Patient with a right ankle injury and delayed presentation. Scannogram was obtained to evaluate the amount of leg length discrepancy
Fig. 1.4
( a–c. Computed tomography scan images showing how to measure malrotation of left tibia—24° internal rotation compared to right side
When looking at the plain radiographs, the absence of bridging bone or callous at the fracture site, sclerotic fracture edges, bone resorption, or persistent fracture lines all may indicate a nonunion. It is imperative to also critically assess the implants and the initial fixation strategy to insure that the original type of healing wanted—primary versus secondary—was being achieved. Often times absolute stability was desired, yet there is callus formation on the radiographs (Fig. 1.5). This may indicate either excessive motion suggesting hardware failure or that the fixation was not as rigid as one wanted, allowing sufficient motion for callous formation. The fracture however may go on to heal. In other situations, it may go on to a nonunion with or without hardware loosening or breakage. Additionally, the radiographs should be assessed for periosteal reaction, loosening/lysis around hardware, and broken implants. Comparison to previous radiographs cannot be overemphasized.
Fig. 1.5
a, b. Injury radiographs (anteroposterior [AP] and lateral) of patient with left humerus fracture after a motor vehicle collision. c, d Postoperative radiographs (AP and lateral) after open reduction and internal fixation performed in an effort to obtain absolute stability. e, f Follow-up radiographs (AP and lateral) showing unintended callus formation due to micromotion despite attempt at rigid fixation—infection workup was negative and patient went on to consolidate
As mentioned above, plain radiography alone is often times not a reliable tool for assessing fracture healing due to the lack of consistency among surgeons and interpretation of the films. It is clear that better ways of assessing fracture healing are needed [6]. Several clinical trials all have shown poor agreement between surgeons [15, 56]. Many have proposed criteria to standardize fracture healing assessment [57, 58]. One such assessment tool is the Radiographic Union Scale for Tibial (RUST) developed by Koolstra and his colleagues [58]. This scoring system assesses the presence or absence of fracture callus and the visibility of the fracture line on each of the four cortices. The scale is from 1 to 3 and based on callus and fracture line visibility at each cortex. A one is the absence of callus and a visible fracture line. A two is the presence of callus, but the fracture line is still visible. A three is for callus and the absence of a fracture line. The minimum score is 4, and the maximum is 12. This has shown to improve agreement for assessing union only in tibia fractures treated with intramedullary nails. Whelan et al. [59] showed an overall inter-observer reliability of 86% and intra-observer reliability of 88%. The RUST score has not been correlated with functional outcomes to date.
A similar scoring system was developed by Bhandari et al. [57] for use in hip fractures. The Radiographic Union Score in Hip Fractures or RUSH was developed to improve agreement in the assessment of femoral neck fractures. In a similar manner to the RUST, the RUSH evaluates cortical bridging on each of the four cortices as well as disappearance of the fracture line and an independent score is given. A one is given for no cortical bridging, two for some cortical bridging, and a three for complete cortical bridging. If the fracture line is visible, a one is given, a two for some evidence of the fracture line, and a three for no evidence of the fracture line. Two other aspects of femoral neck fractures are scored, the trabecular index based on consolidation and the disappearance of the fracture line. A score of 1–3 is assigned as well to each component. The overall minimum is 10 and maximum is 30. Their initial study showed that the RUSH improved agreement among reviewers regardless of subspecialty, but their agreement did not improve over time. A very important shortcoming was that the reviewer’s assessment was found to be potentially inaccurate without information regarding the time of the radiograph. They had 6 of 7 patients deemed as being healed at 2 weeks, which is not possible. Chiavaras et al. [60] extended the RUSH score to evaluate intertrochanteric hip fractures and evaluate agreement between radiologists and orthopedic surgeons. They found that the RUSH score did improve the overall agreement regarding fracture healing from fair to substantial between the two specialties.
Although scoring systems can be beneficial in determining union and providing a more objective measurement over time, the real issue is their use in predicting a nonunion. Recently, Frank et al. [19] did a study to assess the utility of the RUSH score to help define femoral neck fracture nonunion . They retrospectively pulled 250 cases from the FAITH hip fracture trial all of which had 6-month hip radiographs. They determined the RUSH score at 6 months for each case. They found that if the RUSH score at 6 months was <18, it had 100% specificity and a positive predictive value of 100% for a nonunion. They all had a 10 times greater risk of undergoing reoperation for a nonunion. If the patient does develop a nonunion of the femoral neck, a valgus intertrochanteric osteotomy is an option to obtain union. Varghese et al. [61] evaluated a group of 40 patients who underwent the procedure for a femoral neck nonunion developing after neglected fractures. They evaluated the presenting nonunion film for a radiographic index they called the neck resorption ratio (NRR) to determine whether that could predict nonunion of the valgus intertrochanteric osteotomy. The NRR is determined by measuring the length of the fractured head and neck fragment and comparing it to the length of the intact neck on the contralateral side (measured from the tip of the head to the intertrochanteric line). The NRR was found to be the most important factor in predicting union in their series. All patients that had a pre-op NRR of >0.52 had union. Taking this parameter into consideration before making treatment decisions in femoral neck nonunions may allow one to consider a more definitive treatment and avoid a repeat nonunion situation.
Although utilizing a score to predict nonunion after a reconstructive procedure can be useful, a score to predict nonunions for acute fractures would have greater applicability. The Nonunion Risk Determination (NURD) Score was developed by O’Halloran et al. [26]. The authors retrospectively reviewed all tibial shaft fractures at their institution over a 7-year period treated with an intramedullary nail. They had 382 patients with 56 nonunions. Factors were evaluated and they developed a logistic regression model to include seven of these factors. They assigned points to these seven factors. The NURD score gave 1 point for male gender, 2 points for open fractures, 3 points for chronic conditions, 4 points for compartment syndrome, and 5 points for flaps. Additionally, 1 point per ASA grade was given as well as for each 25% reduction of cortical contact (100% = 0; 75% = 1; 50% = 2; 25% = 3). If the injury was low energy or spiral, one point was subtracted for each factor. They found that a NURD score of 0–5 had a 2% chance of nonunion versus a 61% chance if the score was >12. The score was felt to be a potential nonunion prediction model that clinicians could utilize to determine which patients had a higher risk of nonunion. If such scores could be developed and validated for other bones, prediction of nonunions could be commonplace and allow for earlier intervention.
In our practice, comparative plain radiographs over time, and the clinical picture and evaluation of the patient are sufficient to diagnose a nonunion. However, in some situations, plain radiographs may not allow complete evaluation of the nonunion site because of the hardware. In these cases, a CT scan, with metal suppression if hardware is present, can be obtained to further evaluate the nonunion site as well as look for areas of sequestered dead bone or areas of bone deficits that may require bone grafting. CT scans have been shown to have high sensitivity but moderate specificity with about a 90% accuracy for the detection of nonunions [62]. Sagittal and coronal reconstructions to include 3D reconstructions can help with visualization (Fig. 1.6). Many times, these fractures are “clinically” healed, but patients have symptomatic hardware. A CT scan can also aid in looking at the integrity of the bony consolidation and for defects within the “healed” construct. In some situations, the patient can be considered as having an implant dependent union; e.g., there is some central bone loss but sufficient bridging that the bone has healed around these deficiencies, but the strength of the bone may be reliant upon the associated hardware.
Fig. 1.6
Patient referred for nonunion 9 months after treatment for tibial plateau and tibia shaft fracture treated with open reduction and internal fixation. a, b Anteroposterior (AP) and lateral radiographs of the tibia show consolidation of the plateau component. There is hardware failure and nonunion of the tibial shaft. c–f Computed tomography scan images (axial, coronal, sagittal, and 3D reconstructions) which show the subtle hypertrophic nature of the nonunion. g, h One year after treatment of hypertrophic nonunion with hardware removal and subsequent reamed nailing (AP and lateral)
Ultrasonography (US) has been shown to have some utility in diagnosing nonunions [63]. In a study by Moed et al. [64] in which tibia fractures treated with an intramedullary nail were evaluated, the authors showed a sensitivity of 100% and a positive predictive value of 97% in detecting healing of the fracture site. They also could predict healing of these injuries much earlier (38 days versus 127 days) than plain radiography. Chachan et al. [14] in their prospective diagnostic follow-up study showed that ultrasound was able to predict fracture healing 2 weeks earlier than plain radiographs. More importantly, it was able to predict nonunions 8.5 weeks earlier. Despite the earlier detection for a nonunion, US has not become widespread in its use. The benefits of no radiation have not outweighed the primary issues of user dependency, time required for the study and additional cost. Three-dimensional ultrasound is a newer technology that may have added benefits of being able to measure the vascularity not only in the surrounding soft tissue but the fracture itself, as well as providing more information on the progression of healing [63].
Fluoroscopy is another imaging modality that can be used primarily to assess motion at a fracture site to determine healing. This is most useful in the patient treated without internal fixation and when there is a question of the healed status of the injury. It can also be useful in cases where external fixation has been used since the external fixation can be loosened without complete removal and the fracture site stressed. If there is motion, the external fixation can easily be “tightened” and “reset.” In our practice, this is usually done in conjunction with anticipated external fixation removal after definitive management of a fracture or in reconstructive cases, where determining the “laxity” of a nonunion can guide treatment.
Magnetic resonance imaging (MRI) with gadolinium can be used to assess the nonunion site for infection and more importantly for devascularized bone or a sequestrum [31, 65]. Additionally, because of its ability to detect marrow changes, it is very sensitive for osteomyelitis. Osteomyelitis usually shows decreased marrow signal on T-1 images but increased signal on T-2 images (Fig. 1.7a–d). The MRI also allows one to determine the extent of bony involvement [31] in such cases because of the marrow changes which is crucial in determining the best reconstructive option based upon the anticipated length of resection required to eradicate the osteomyelitis.
Fig. 1.7
a, b Anteroposterior and lateral radiographs of patient with infected nonunion of right tibia. c, d Magnetic resonnance images of tibia showing increased signal on T2 image indicating osteomyelitis. e Nuclear medicine studies showing increased uptake on indium study suggesting the presence of infection
Nuclear medicine studies (Fig. 1.7e) have been historically used to aid in the detection of infection as well, but over time their utility has been questioned [66]. They are still of use in evaluating the nonunion site for infection and/or biologic activity [31]. Leukocyte-labeled studies have been shown to have appropriate diagnostic accuracy for osteomyelitis in the peripheral skeleton [67]. The traditional technetium bone scan will have increased signal on any biological bone activity, and hence, any fracture site that is biologically active should have uptake. Thus, it really is not used for the evaluation of healing although, in cases of avascular or nonviable fractures sites, e.g., the atrophic nonunion , decreased or no uptake may be the case. Our use is usually for the suspected infected cases when the clinical signs of an infection are absent but laboratory markers—erythrocyte sedimentation rate (ESR) , C-reactive protein (CRP) , or white blood cell (WBC) count —may be elevated. In these cases, a bone scan is obtained which is usually positive. If by chance it is negative for uptake, then no other imaging is done and concern becomes for an atrophic nonunion. After a positive bone scan, an indium (tagged WBC) scan is performed. If this is positive at the fracture/nonunion site, then there is increased suspicion for infection. If it is negative for uptake at the site, then infection is less likely but unfortunately never completely ruled out. The final study done after a positive Indium scan, is the sulfa colloid marrow scan. The areas of uptake are then compared to the indium scan. If the areas of uptake are concordant with the indium scan , the uptake is deemed to be secondary to the associated marrow changes and not infection. Conversely, if the areas of uptake on the indium scan do not coincide with uptake on the sulfa colloid (discordant), then it is thought to be suggestive of an infection [31]. The specificity and sensitivity of such imaging studies has been controversial. Stucken et al. [66] showed that not utilizing the nuclear medicine tests actually improved their predicted probabilities of infection based on laboratory studies alone. The latest imaging modality, which has shown some promise to aid in the detection of infection or osteomyelitis, has been the positron emission tomography (PET) scan ± CT scan. A fluorodeoxyglucose PET scan has been shown to have the highest diagnostic accuracy for excluding or confirming the diagnosis of chronic osteomyelitis [31, 67]. This could aid in the evaluation of the presence of infection in a nonunion.
1.4.3 Laboratory Evaluation
Laboratory studies can assist in determining the etiology of the nonunion or at least look at conditions that may have contributed to the development of the nonunion. All patients should be evaluated with a CBC with differential, ESR, and CRP. These are utilized to evaluate for infection but realizing that the ESR and CRP are simply indicators of inflammation and can be elevated in the absence of an infected nonunion. Conversely, normal markers do not necessarily rule out an infection either and are usually the case in indolent infections. A standardized protocol to rule out infection was assessed by Stucken et al. [66] to evaluate the efficacy of laboratory studies (WBC, CRP, ESR) and nuclear medicine studies. They found that the ESR and the CRP were both independently accurate predictors of infection. With all three tests being positive, the predicted probability of an infection was 100%. If the nuclear medicine studies were included, the probability went down to 86% for three positive tests.
As mentioned before, in cases where the original surgery was deemed to be highly contributory to the development of the nonunion, more extensive laboratory studies may not be needed. In the cases where the technical aspects seemed to be sound and the reason for the nonunion unclear, other laboratory studies may point to an underlying metabolic abnormality as the etiology [13]. These patients would probably benefit from an endocrinology workup if feasible. Often times in our practice, these are unfunded trauma patients and the workup is often left to the orthopedic trauma surgeon to do the full evaluation. Many of these patients may also have sustained fragility fractures which also warrant laboratory workup. These underlying metabolic disorders include vitamin D deficiency, hypothyroidism, hypogonadism, hypocalcemia, and overall poor nutritional status. Brinker et al. [13] showed that 31 of 37 of their patients with a nonunion had some type of metabolic abnormality with vitamin D deficiency being the most common. The laboratory studies, in addition to the above, should include serum 25-hydroxy-vitamin D, calcium, phosphorus, alkaline phosphatase, thyroid function tests, parathyroid hormone level, hormone levels (testosterone, estrogen, and follicle stimulating hormone), and albumin and cortisol levels [13]. Vitamin D deficiency has been set at 20–30 mg/dl, and <20 are considered insufficient.
In cases of infected nonunions, it is helpful to obtain results of previous cultures if available to determine the previous organism(s). At the time of surgery, especially in cases of staged procedures, which is often the situation in dealing with infected nonunions [66], deep tissue cultures and bone biopsies can help determine the presence or absence of an infection as well as the offending organism. Preoperative antibiotics should be withheld until after intra-operative cultures are obtained. It is also recommended to cease any antibiotics for at least two weeks, if possible, to maximize the chance of identifying the organism. The first stage is usually to remove previous hardware, to obtain a better idea of the nonunion site, and to get biopsies and cultures. Antibiotic beads can be placed in the interim prior to the second stage.
There has been an increasing interest in looking for serologic markers that may help to predict fracture healing and therefore potentially predict nonunions [16, 20, 22, 27, 30, 68]. Although a full review and discussion of these markers is beyond the scope of this chapter, it is important to mention that they exist and have future implications in predicting fracture healing. These biomarkers are either factors that regulate the healing process itself or bone turnover markers that are extracellular matrix components related to degradation or production during the repair process [15]. The local or systemic factors regulating the healing process include vascular endothelial growth factor (VEGF) and transforming growth factor-beta (TGF-β) . Serum TGF-β has been found to be an indicator of healing versus nonhealing with 4 week levels being much lower in a group of patients that had a delayed union [20]. The bone turnover markers can be divided into one of three categories: 1. bone formation markers, 2. bone resorption markers, and 3. osteoclast regulatory proteins [16, 20, 22]. The bone formation markers indicate osteoblastic activity and as such are fragments of type-I and type-III pro-collagen that are released during the formation of type-III collagen (PIIINP, PICP, PIIINP). Osteocalcin (OC) and bone-specific alkaline phosphatase (BSAP) are also measures of osteoblastic activity. Bone resorption markers include those that measure the degradation of type-I collagen (CTX, NTX, ICTP, pyridinoline, deoxypyridinoline). Tartrate-resistant acid phosphatase (TRAcP) and cathepsin K (CK) are noncollagenous markers that also measure bone resorption but are osteoclast regulatory proteins. Other osteoclast regulatory proteins include receptor activator of nuclear factor NF-kB ligand (RANKL) and osteoprotegerin (OPG). The marker activity of only a handful of these have been evaluated in various fractures and shown some promise in predicting fracture healing [30]. Fischer et al. [18] evaluated a number of cytokines—TGF-β, platelet-derived growth factor (PDGF-AB), insulin-like growth factor 1 (IGF-1)—in patients with long-bone nonunions treated both successfully and unsuccessfully with the Masquelet technique and compared them to a group with normal bone healing. They found temporal variations of these cytokines in the three groups, with high expressions of IGF-1 corresponding to a successful Masquelet treatment. They demonstrated significant differences in cytokine expression between normal fracture healing and the nonunion treatment groups. If the time profiles of each of these markers can be fully understood, then perhaps variations in these markers from what may be considered the normal in fracture healing may provide insight into which fractures will go on to a nonunion [27]. Earlier detection and subsequent earlier treatment could result in substantial cost savings [20].
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
